Skip to contentSkip to conversation

PM / devfreq: try_then_request_governor should not return NULL

PATCH In review Rob Clark 3 participants 4 messages May 23, 2019 – May 28, 2019

Subsystems (2)
lore
Share & tools
lore

Rob Clark

author
#0 lore raw

PATCH 1/1 PM / devfreq: try_then_request_governor should not return NULL

From: Rob Clark <[email protected]>

The two spots it is called expect either an IS_ERR() or a valid pointer, but not NULL.

Fixes this crash that I came across:

   Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030
   Mem abort info:
     ESR = 0x96000005
     Exception class = DABT (current EL), IL = 32 bits
     SET = 0, FnV = 0
     EA = 0, S1PTW = 0
   Data abort info:
     ISV = 0, ISS = 0x00000005
     CM = 0, WnR = 0
   [0000000000000030] user address but active_mm is swapper
   Internal error: Oops: 96000005 [#1] PREEMPT SMP
   Modules linked in:
   Process kworker/2:1 (pid: 212, stack limit = 0x(____ptrval____))
   CPU: 2 PID: 212 Comm: kworker/2:1 Not tainted 5.1.0-43338-g460e6984675c-dirty #54
   Hardware name: Google Cheza (rev3+) (DT)
   Workqueue: events deferred_probe_work_func
   pstate: 00c00009 (nzcv daif +PAN +UAO)
   pc : devfreq_add_device+0x2e4/0x410
   lr : devfreq_add_device+0x2d4/0x410
   sp : ffffff8013d93740
   x29: ffffff8013d93790 x28: ffffffc0f54f8670
   x27: 0000000000000001 x26: 0000000000000007
   x25: ffffff80124abfd8 x24: 0000000000000000
   x23: ffffffc0fabc4048 x22: ffffffc0fabc4388
   x21: ffffffc0fabc4010 x20: ffffffc0fa243010
   x19: ffffffc0fabc4000 x18: 0000000091c3d373
   x17: 0000000000000400 x16: 000000000000001a
   x15: 000000019e06d400 x14: 0000000000000001
   x13: 0000000000000000 x12: 00000000000006b6
   x11: 0000000000000000 x10: 0000000000000000
   x9 : ffffffc0fa18ba00 x8 : 0000000000000000
   x7 : 0000000000000000 x6 : ffffff80127a3d9a
   x5 : ffffff8013d93550 x4 : 0000000000000000
   x3 : 0000000000000000 x2 : 0000000000000000
   x1 : 00000000000000fe x0 : 0000000000000000
   Call trace:
    devfreq_add_device+0x2e4/0x410
    devm_devfreq_add_device+0x64/0xac
    msm_gpu_init+0x320/0x5c0
    adreno_gpu_init+0x21c/0x274
    a6xx_gpu_init+0x68/0xf4
    adreno_bind+0x158/0x284
    component_bind_all+0x110/0x204
    msm_drm_bind+0x118/0x5b8
    try_to_bring_up_master+0x15c/0x19c
    component_master_add_with_match+0xb4/0xec
    msm_pdev_probe+0x1f0/0x27c
    platform_drv_probe+0x90/0xb0
    really_probe+0x120/0x298
    driver_probe_device+0x64/0xfc
    __device_attach_driver+0x8c/0xa4
    bus_for_each_drv+0x88/0xd0
    __device_attach+0xac/0x134
    device_initial_probe+0x20/0x2c
    bus_probe_device+0x34/0x90
    deferred_probe_work_func+0x74/0xac
    process_one_work+0x210/0x428
    worker_thread+0x278/0x3e4
    kthread+0x120/0x130
    ret_from_fork+0x10/0x18
   Code: aa0003f8 b13ffc1f 54000762 f901c278 (f9401b08)
   ---[ end trace a6ecc18ce5894375 ]---
   Kernel panic - not syncing: Fatal exception
---
 drivers/devfreq/devfreq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/devfreq/devfreq.c +1 -1
index 0ae3de76833b..d29f66f0e52a 100644
@@ -254,7 +254,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
 		/* Restore previous state before return */
 		mutex_lock(&devfreq_list_lock);
 		if (err)
-			return NULL;
+			return ERR_PTR(err);
 
 		governor = find_devfreq_governor(name);
 	}
Signature · 1 line
2.20.1
maintainer
#1 lore raw

Hi,

This issue[1] is already fixed on latest linux.git You can check it. Thanks.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b53b0128052ffd687797d5f4deeb76327e7b5711

Regards,
Chanwoo Choi

Quoted 90 lines · Rob Clark

From: Rob Clark <[email protected]>

The two spots it is called expect either an IS_ERR() or a valid pointer, but not NULL.

Fixes this crash that I came across:

   Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030
   Mem abort info:
     ESR = 0x96000005
     Exception class = DABT (current EL), IL = 32 bits
     SET = 0, FnV = 0
     EA = 0, S1PTW = 0
   Data abort info:
     ISV = 0, ISS = 0x00000005
     CM = 0, WnR = 0
   [0000000000000030] user address but active_mm is swapper
   Internal error: Oops: 96000005 [#1] PREEMPT SMP
   Modules linked in:
   Process kworker/2:1 (pid: 212, stack limit = 0x(____ptrval____))
   CPU: 2 PID: 212 Comm: kworker/2:1 Not tainted 5.1.0-43338-g460e6984675c-dirty #54
   Hardware name: Google Cheza (rev3+) (DT)
   Workqueue: events deferred_probe_work_func
   pstate: 00c00009 (nzcv daif +PAN +UAO)
   pc : devfreq_add_device+0x2e4/0x410
   lr : devfreq_add_device+0x2d4/0x410
   sp : ffffff8013d93740
   x29: ffffff8013d93790 x28: ffffffc0f54f8670
   x27: 0000000000000001 x26: 0000000000000007
   x25: ffffff80124abfd8 x24: 0000000000000000
   x23: ffffffc0fabc4048 x22: ffffffc0fabc4388
   x21: ffffffc0fabc4010 x20: ffffffc0fa243010
   x19: ffffffc0fabc4000 x18: 0000000091c3d373
   x17: 0000000000000400 x16: 000000000000001a
   x15: 000000019e06d400 x14: 0000000000000001
   x13: 0000000000000000 x12: 00000000000006b6
   x11: 0000000000000000 x10: 0000000000000000
   x9 : ffffffc0fa18ba00 x8 : 0000000000000000
   x7 : 0000000000000000 x6 : ffffff80127a3d9a
   x5 : ffffff8013d93550 x4 : 0000000000000000
   x3 : 0000000000000000 x2 : 0000000000000000
   x1 : 00000000000000fe x0 : 0000000000000000
   Call trace:
    devfreq_add_device+0x2e4/0x410
    devm_devfreq_add_device+0x64/0xac
    msm_gpu_init+0x320/0x5c0
    adreno_gpu_init+0x21c/0x274
    a6xx_gpu_init+0x68/0xf4
    adreno_bind+0x158/0x284
    component_bind_all+0x110/0x204
    msm_drm_bind+0x118/0x5b8
    try_to_bring_up_master+0x15c/0x19c
    component_master_add_with_match+0xb4/0xec
    msm_pdev_probe+0x1f0/0x27c
    platform_drv_probe+0x90/0xb0
    really_probe+0x120/0x298
    driver_probe_device+0x64/0xfc
    __device_attach_driver+0x8c/0xa4
    bus_for_each_drv+0x88/0xd0
    __device_attach+0xac/0x134
    device_initial_probe+0x20/0x2c
    bus_probe_device+0x34/0x90
    deferred_probe_work_func+0x74/0xac
    process_one_work+0x210/0x428
    worker_thread+0x278/0x3e4
    kthread+0x120/0x130
    ret_from_fork+0x10/0x18
   Code: aa0003f8 b13ffc1f 54000762 f901c278 (f9401b08)
   ---[ end trace a6ecc18ce5894375 ]---
   Kernel panic - not syncing: Fatal exception
---
 drivers/devfreq/devfreq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/devfreq/devfreq.c +1 -1
index 0ae3de76833b..d29f66f0e52a 100644
@@ -254,7 +254,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
 		/* Restore previous state before return */
 		mutex_lock(&devfreq_list_lock);
 		if (err)
-			return NULL;
+			return ERR_PTR(err);
 
 		governor = find_devfreq_governor(name);
 	}
Signature · 3 lines
Best Regards,
Chanwoo Choi
Samsung Electronics
#2 lore raw

Ahh, thanks, I've not moved to the latest -rc yet..

That commit would be a good candidate for 5.1.y stable branch

BR,
-R

Quoted 109 lines · Chanwoo Choi

Hi,

This issue[1] is already fixed on latest linux.git You can check it. Thanks.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b53b0128052ffd687797d5f4deeb76327e7b5711

Regards,
Chanwoo Choi

Quoted 82 lines · Rob Clark

From: Rob Clark <[email protected]>

The two spots it is called expect either an IS_ERR() or a valid pointer, but not NULL.

Fixes this crash that I came across:

   Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030
   Mem abort info:
     ESR = 0x96000005
     Exception class = DABT (current EL), IL = 32 bits
     SET = 0, FnV = 0
     EA = 0, S1PTW = 0
   Data abort info:
     ISV = 0, ISS = 0x00000005
     CM = 0, WnR = 0
   [0000000000000030] user address but active_mm is swapper
   Internal error: Oops: 96000005 [#1] PREEMPT SMP
   Modules linked in:
   Process kworker/2:1 (pid: 212, stack limit = 0x(____ptrval____))
   CPU: 2 PID: 212 Comm: kworker/2:1 Not tainted 5.1.0-43338-g460e6984675c-dirty #54
   Hardware name: Google Cheza (rev3+) (DT)
   Workqueue: events deferred_probe_work_func
   pstate: 00c00009 (nzcv daif +PAN +UAO)
   pc : devfreq_add_device+0x2e4/0x410
   lr : devfreq_add_device+0x2d4/0x410
   sp : ffffff8013d93740
   x29: ffffff8013d93790 x28: ffffffc0f54f8670
   x27: 0000000000000001 x26: 0000000000000007
   x25: ffffff80124abfd8 x24: 0000000000000000
   x23: ffffffc0fabc4048 x22: ffffffc0fabc4388
   x21: ffffffc0fabc4010 x20: ffffffc0fa243010
   x19: ffffffc0fabc4000 x18: 0000000091c3d373
   x17: 0000000000000400 x16: 000000000000001a
   x15: 000000019e06d400 x14: 0000000000000001
   x13: 0000000000000000 x12: 00000000000006b6
   x11: 0000000000000000 x10: 0000000000000000
   x9 : ffffffc0fa18ba00 x8 : 0000000000000000
   x7 : 0000000000000000 x6 : ffffff80127a3d9a
   x5 : ffffff8013d93550 x4 : 0000000000000000
   x3 : 0000000000000000 x2 : 0000000000000000
   x1 : 00000000000000fe x0 : 0000000000000000
   Call trace:
    devfreq_add_device+0x2e4/0x410
    devm_devfreq_add_device+0x64/0xac
    msm_gpu_init+0x320/0x5c0
    adreno_gpu_init+0x21c/0x274
    a6xx_gpu_init+0x68/0xf4
    adreno_bind+0x158/0x284
    component_bind_all+0x110/0x204
    msm_drm_bind+0x118/0x5b8
    try_to_bring_up_master+0x15c/0x19c
    component_master_add_with_match+0xb4/0xec
    msm_pdev_probe+0x1f0/0x27c
    platform_drv_probe+0x90/0xb0
    really_probe+0x120/0x298
    driver_probe_device+0x64/0xfc
    __device_attach_driver+0x8c/0xa4
    bus_for_each_drv+0x88/0xd0
    __device_attach+0xac/0x134
    device_initial_probe+0x20/0x2c
    bus_probe_device+0x34/0x90
    deferred_probe_work_func+0x74/0xac
    process_one_work+0x210/0x428
    worker_thread+0x278/0x3e4
    kthread+0x120/0x130
    ret_from_fork+0x10/0x18
   Code: aa0003f8 b13ffc1f 54000762 f901c278 (f9401b08)
   ---[ end trace a6ecc18ce5894375 ]---
   Kernel panic - not syncing: Fatal exception
---
 drivers/devfreq/devfreq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/devfreq/devfreq.c +0 -0
index 0ae3de76833b..d29f66f0e52a 100644

drivers/devfreq/devfreq.c static struct devfreq_governor *try_then_request_governor(const char *name) · line 260

Quoted 8 lines
drivers/devfreq/devfreq.c +1 -1
@@ -254,7 +254,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
              /* Restore previous state before return */
              mutex_lock(&devfreq_list_lock);
              if (err)
-                     return NULL;
+                     return ERR_PTR(err);
 
              governor = find_devfreq_governor(name);
      }

--
Best Regards,
Chanwoo Choi
Samsung Electronics

maintainer
#3 lore raw

Hi Rob,

You're right. It have to be posted to [email protected]. As you recommended, I send it[1] to [email protected] for fixup. [1]https://lkml.org/lkml/2019/5/27/547

Best Regards,
Chanwoo Choi

Quoted 119 lines · Rob Clark

Ahh, thanks, I've not moved to the latest -rc yet..

That commit would be a good candidate for 5.1.y stable branch

BR,
-R

Quoted 109 lines · Chanwoo Choi

Hi,

This issue[1] is already fixed on latest linux.git You can check it. Thanks.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=b53b0128052ffd687797d5f4deeb76327e7b5711

Regards,
Chanwoo Choi

Quoted 82 lines · Rob Clark

From: Rob Clark <[email protected]>

The two spots it is called expect either an IS_ERR() or a valid pointer, but not NULL.

Fixes this crash that I came across:

   Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030
   Mem abort info:
     ESR = 0x96000005
     Exception class = DABT (current EL), IL = 32 bits
     SET = 0, FnV = 0
     EA = 0, S1PTW = 0
   Data abort info:
     ISV = 0, ISS = 0x00000005
     CM = 0, WnR = 0
   [0000000000000030] user address but active_mm is swapper
   Internal error: Oops: 96000005 [#1] PREEMPT SMP
   Modules linked in:
   Process kworker/2:1 (pid: 212, stack limit = 0x(____ptrval____))
   CPU: 2 PID: 212 Comm: kworker/2:1 Not tainted 5.1.0-43338-g460e6984675c-dirty #54
   Hardware name: Google Cheza (rev3+) (DT)
   Workqueue: events deferred_probe_work_func
   pstate: 00c00009 (nzcv daif +PAN +UAO)
   pc : devfreq_add_device+0x2e4/0x410
   lr : devfreq_add_device+0x2d4/0x410
   sp : ffffff8013d93740
   x29: ffffff8013d93790 x28: ffffffc0f54f8670
   x27: 0000000000000001 x26: 0000000000000007
   x25: ffffff80124abfd8 x24: 0000000000000000
   x23: ffffffc0fabc4048 x22: ffffffc0fabc4388
   x21: ffffffc0fabc4010 x20: ffffffc0fa243010
   x19: ffffffc0fabc4000 x18: 0000000091c3d373
   x17: 0000000000000400 x16: 000000000000001a
   x15: 000000019e06d400 x14: 0000000000000001
   x13: 0000000000000000 x12: 00000000000006b6
   x11: 0000000000000000 x10: 0000000000000000
   x9 : ffffffc0fa18ba00 x8 : 0000000000000000
   x7 : 0000000000000000 x6 : ffffff80127a3d9a
   x5 : ffffff8013d93550 x4 : 0000000000000000
   x3 : 0000000000000000 x2 : 0000000000000000
   x1 : 00000000000000fe x0 : 0000000000000000
   Call trace:
    devfreq_add_device+0x2e4/0x410
    devm_devfreq_add_device+0x64/0xac
    msm_gpu_init+0x320/0x5c0
    adreno_gpu_init+0x21c/0x274
    a6xx_gpu_init+0x68/0xf4
    adreno_bind+0x158/0x284
    component_bind_all+0x110/0x204
    msm_drm_bind+0x118/0x5b8
    try_to_bring_up_master+0x15c/0x19c
    component_master_add_with_match+0xb4/0xec
    msm_pdev_probe+0x1f0/0x27c
    platform_drv_probe+0x90/0xb0
    really_probe+0x120/0x298
    driver_probe_device+0x64/0xfc
    __device_attach_driver+0x8c/0xa4
    bus_for_each_drv+0x88/0xd0
    __device_attach+0xac/0x134
    device_initial_probe+0x20/0x2c
    bus_probe_device+0x34/0x90
    deferred_probe_work_func+0x74/0xac
    process_one_work+0x210/0x428
    worker_thread+0x278/0x3e4
    kthread+0x120/0x130
    ret_from_fork+0x10/0x18
   Code: aa0003f8 b13ffc1f 54000762 f901c278 (f9401b08)
   ---[ end trace a6ecc18ce5894375 ]---
   Kernel panic - not syncing: Fatal exception
---
 drivers/devfreq/devfreq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/devfreq/devfreq.c +0 -0
index 0ae3de76833b..d29f66f0e52a 100644

drivers/devfreq/devfreq.c static struct devfreq_governor *try_then_request_governor(const char *name) · line 260

Quoted 8 lines
drivers/devfreq/devfreq.c +1 -1
@@ -254,7 +254,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
              /* Restore previous state before return */
              mutex_lock(&devfreq_list_lock);
              if (err)
-                     return NULL;
+                     return ERR_PTR(err);
 
              governor = find_devfreq_governor(name);
      }

--
Best Regards,
Chanwoo Choi
Samsung Electronics

Signature · 3 lines
Best Regards,
Chanwoo Choi
Samsung Electronics